Abstract Basal-like breast cancer (BLBC) usually does not express ER, PR, and HER2. It is among the most aggres- sive forms of breast cancer, and yet there is no targeted therapy to treat it because we do not know the drivers for this disease. The cell of origin of BLBC is also mysterious. While up to 20% of invasive breast cancers are basal-like, basal-like lesions are undetectable in very early stages of premalignancy, such as atypical ductal hyperplasia (ADH) ? these early premalignant lesions are nearly all luminal. Only at a later premalignant stage, ductal carcinoma in situ (DCIS), can basal-like lesions begin to be seen in ?5% of the cases. These ob- servations support the possibility that BLBC evolves from cells with luminal properties upon activation of a driver pathway during DCIS. The overall objective of this exploratory project is to define such a driver that can promote the development of BLBC, leading to a treatment strategy that can prevent BLBC early in its tracks. We have recently reported that N-Ras may be such a driver. While oncogenic RAS mutations are rare in breast cancer, wild-type RAS genes are differentially overexpressed in different breast cancer subtypes ? wild-type N-RAS is overexpressed in BLBCs. We further showed that high expression of N-RAS, as well as of those genes whose expression is N-Ras-dependent (encoding mostly cyto/chemokines, such as IL8), corre- lates with poor patient outcome. The overexpressed N-Ras is both required and sufficient for promoting tumor- igenesis, via stimulating the JAK2-STAT5-IL8 pathway rather than the conventional MAP kinase pathway. In follow-up and on-going studies, we found that activation of this N-Ras pathway correlated with basal-like mark- ers in DCIS patients, and N-Ras overexpression in an established luminal DCIS cell line induced acquisition of basal-like properties in vitro. To properly study DCIS in vivo, we have pioneered the technique of injecting DCIS cells intraductally into mouse mammary glands, allowing human DCISs to develop in an appropriate tis- sue environment to closely mimic human DCIS progression. When such N-Ras-overexpressing luminal cells were so transplanted into mice, basal-like markers were also detected. We thus hypothesize that BLBC evolves from luminal cells during DCIS, driven by N-Ras. To firm up this mechanism and to explore treatment/ prevention options for BLBC, in Aim 1 we will assess whether N-Ras activity at DCIS can promote a luminal to basal-like switch and increase invasiveness, by overexpressing it in luminal cell lines and patient-derived xeno- graft models. Conversely, we will silence N-RAS expression in basal-like DCIS cells to determine whether the converse can be seen. In Aim 2 we will assess whether formation of BLBC can be blocked early during DCIS by pharmacologically inhibiting the JAK2-STAT5 pathway downstream of N-Ras in xenograft models. In addi- tion to FDA-approved drugs against JAK2, we will also examine a novel STAT5 inhibitor that will soon enter clinical trials. If N-Ras indeed promotes a switch from luminal to BLBC during DCIS, blocking its downstream effectors could prevent development of the most aggressive form of breast cancer, BLBC.